Research Of Lane Departure Warning Method Based On Kalman Filter Tracking

With the rapid increasing number of car ownership, the rate of traffic accidents increases sharply, in which the proportion caused by lane departure reached 30%. The safety of traffic and driving has become a social hot spot and attracted people’s attention. In order to reduce the traffic accidents the government and researchers have paid much attention on the lane departure warning. By utilized lane departure warning system it can be reduced at least 24% accidents induced by lane departure. Therefore lane departure warning researching has important application significance in traffic safety improvement, traffic accident reduction and avoidance of the loss caused by traffic accidents.Depending on the review on research and application of domestic and foreign lane departure warning study, and also considering of the actual situation of highway in our country, it is proposed a lane departure warning system solution based on Hough transform and Kalman filter.The whole system consists of four modules, i.e., image preprocessing, lane detection, lane tracking and departure warning. In order to reduce the interference and to narrow detection scope, the preprocessing module processes the video image by frame via image graying, interest region selection, and binarization. Finally it gives out the truncated binary images for the subsequent module. In detection module, it first assumes a linear model for the lanes, and then uses Hough transformation to detect the lane lines, finally calculates the lines’ parameters. In the tracking, Kalman filter is explored to track the detected lines based on the assumption that the lines are moving linearly and continuously. In order to reduce noise and improve stability during detection and tracking, the estimated values of matched Kalman filter replace the detected values from Hough transformation. In the last module, the lane departure warning algorithm based on lateral distance is realized.Finally the proposed solution is verified on the actual road videos. It can accurately detect the lanes and in time make departure warnings. The results show that the proposed solution has good early-warning accuracy and real-time calculation. It is an effective solution of lane departure warning system with a certain practical significance.